Building a 3-D World Model for a Mobile Robot from Sensory Data

摘要

This paper presents a method for building a 3-D world model for a mobile robot from sensory data. The 3-D world model consists for three kinds for maps: a sensor map, a local map and a global map. A range image (sensor map) is transformed to a height map (local map) with respect to a mobile robot. First, the height map is segmented into four categories (local map) with respect to a mobile robot. First, the height map is segmented into four categories (unexplored, occluded, traversable,,and obstacle regions) for obstacle detection and path planning. Next, obstacle regions are classified into artificial objects (buildings, cars, and road sings, etc.) or natural objects (trees, bushes, etc.) using both the height image and video image. One drawback of the height map--the recovery of vertical planes--is overcome by the utilization of multiple height maps which include the maximum and minimum height of each point, and the number of points in the range image mapped into one point in the height map. The multiple height map is useful not only for finding vertical planes in the height map but also for segmentation of the video image. Finally, the height maps are integrated into a global map by matching geometrical properties and updating region labels. The method is tested on a model including many objects, such as trees, buildings, cars, and so on.